Magazine of Concrete Research Magazine of Concrete Research

http://dx.doi.org/10.1680/macr.15.00097
Paper 1500097
Jones, Ozlutas and Zheng Received 03/03/2015; revised 30/04/2015; accepted 06/05/2015
Stability and instability of foamed concrete ICE Publishing: All rights reserved

Stability and instability of

foamed concrete
M. Roderick Jones Li Zheng
Professor, Concrete Technology Unit, Division of Civil Engineering, Doctor, Concrete Technology Unit, Division of Civil Engineering,
University of Dundee, Dundee, UK University of Dundee, Dundee, UK
Kezban Ozlutas
Doctor, Concrete Technology Unit, Division of Civil Engineering,
University of Dundee, Dundee, UK

Foamed concrete has proven to be an effective alternative to granular fills and is now widely used internationally.
With increasing demand for lightweight materials for buildings in order to improve sustainability, foamed concrete
has also developed as an ideal material for this purpose, and many countries utilise construction with precast
foamed concrete blocks. However, at densities lower than current technology allows, typically <500 kg/m3, foamed
concretes are more prone to instability of the fresh mix. Furthermore at very low densities, ≤300 kg/m3, instability is
almost inevitable, greatly limiting the potential of foamed concrete for applications where mass is critical (e.g. weak
soils, backfilling damaged structure etc.). This paper aims to illustrate the mechanisms of stability and instability in
foamed concretes and demonstrates how ultra-low density mixes (down to plastic density of 150 kg/m3) can be
successfully produced.

Notation
Fb bubble buoyancy force
Fc bubble confinement force
Fd drainage force
Fst surface tension of bubbles
Φ bubble diameter
Pi internal bubble pressure
r bubble radius
γ interfacial surface tension
Introduction and background
Foamed concrete is now widely used internationally, but
growing pressures for more sustainable construction technol-
ogies, such as lightening of structures, energy conservation, mini-
mising the use of primary resources, and resource efficiency, as
well as reducing the impact of environmental noise, underpin a
need for the development of ultra-low density foamed concrete
that current technology is unable to achieve.
‘Conventional’ foamed concrete can be regarded as having
a plastic density of 500–1600 kg/m3, and ultra-low density
foamed concrete has been defined as having a plastic density
≤500 kg/m3. However, it has previously been reported that
mixes at these latter densities have greatly increased suscepti-
bility to instability (Aldridge, 2005; Jones and McCarthy, 2005,
2006). Indeed, at ≤300 kg/m3, consistently achieving stable
foamed concrete mixes is extremely difficult. Instability of
foamed concrete is segregation of the fresh mix due to the sep-
aration of solids and air phases in the mix. Generally, this
segregation is catastrophic, leading to a complete loss of the air
phase and leaving only the base mix. However, there is no clear
understanding of the underlying mechanism of bubble stability
in foamed concrete mixes, or why ultra-low density mixes are
more prone to becoming unstable.
Working with industry, the authors are aware that this has led
to an inability to deploy ultra-low density foamed concrete,
even though there is a demand from many construction sectors.
Based on laboratory-based studies carried out over a decade,
the authors have attempted to develop an empirical understand-
ing of the factors that have been identified as being critical
to bubble stability, which are reported here, and thereby
develop a method for consistently producing ultra-low density
mixes.
Fundamental issues and observations of
foamed concrete stability
Effect of plastic density on bubble size
Figure 1 illustrates the typical appearance of instability in
foamed concrete mixes, both in a laboratory (Figure 1(a)) and
on site (Figures 1(b) and 1(c)). Figure 1(b) illustrates a tran-
sition point where a mix is becoming unstable, and bubbles
have risen to the surface. This can happen from almost im-
mediately to, more typically, after tens of minutes, but has also
been noted up to 24 h after placement. Observationally, the
lower the plastic density of the mix, the shorter the time to the
onset of instability.
It has been previously noted that foamed concrete average
bubble size increases with decreasing plastic density (Jones

1
Magazine of Concrete Research
(a)
(b)
Figure 1. Examples of instability of ultra-low density foamed
concrete: (a) in the laboratory; (b) and (c) on site
and Zheng, 2013; Nambiar and Ramamurthy, 2007;
She et al., 2014; Visagie and Kearsley, 1999). This is, initially,
a surprising observation as the nature of the input foam is the
same for all mixes and thus bubbles must change size once
combined with the base mix. It is not possible to say whether
this is an immediate or more gradual process, but it does
underline the fact that bubble formation is a dynamic process
rather than a simple incorporation of more or less foam into a
base mix. It is also not clear whether foamed concrete bubbles
2
Jones, Ozlutas and Zheng
Stability and instability of foamed concrete
(c)
are either larger or smaller than the bubble size of the parent
foam, as it is difficult to obtain bubble size metrics in wet
foam. However, once this dynamic stage is complete the
bubbles form a size essentially proportional to the plastic
density of the fresh mix.
Figures 2(a) and 2(b) show typical examples of protein-based
foamed concretes with plastic densities of 1000 kg/m3 and
500 kg/m3 respectively. This shows that a larger bubble size
Magazine of Concrete Research
(a)
(b)
Figure 2. Comparative bubble size in (a) high and (b) low plastic
density foamed concretes
also results in a thinner ‘wall’ separating adjacent bubbles. In
addition, these thinner walls tend to contain many more
‘small’ bubbles. It is speculated that these bubbles are due to
entrapped air in the base mix, which, when constricted in the
thin walls, become more visible. The potential effect of these is
discussed below.
Forces acting on bubbles in fresh foamed concrete
mixes
Figure 3 is an attempt to provide a two-dimensional (2D) sche-
matic illustration of the ‘forces’ acting on a single bubble when
incorporated into fresh foamed concrete. Based on this, a
stable equilibrium state of the bubble is obtained when the
bubble confinement force (Fc), the drainage force (Fd), internal
bubble pressure (Pi), surface tension of the bubbles due to the
effect of surfactant (Fst) and the bubble buoyancy force (Fb)
are balanced.
Jones, Ozlutas and Zheng
Stability and instability of foamed concrete
Fb
Surface Fst
charges
Individual + Φ
cement + Pi Fc
particles +
Fd
Figure 3. Idealised schematic illustration of forces acting on a
single bubble within a fresh foamed concrete mix
The bubble confinement force (Fc) is mainly due to the plastic
density of the fresh mix, but the type of constituent materials,
such as use of different fillers (e.g. sand or coarse fly ash) and
cement type also affect Fc, and can be related to the prevailing
mix rheological characteristics of yield stress and plastic vis-
cosity. At densities ≤500 kg/m3, where sand/filler is generally
not used, there is likely to be a significant decrease in Fc due
to a decrease in yield stress, which results in larger and more
closely spaced bubbles. To reach ultra-low densities, both the
cement and water contents of the mix have to be reduced and
hence Fc is reduced. On the other hand, there is evidence to
support the use of finer cementitious materials (such as fly ash)
to provide enhanced particle packing around the bubbles and a
greater confining force (Nambiar and Ramamurthy, 2007), and
hence smaller bubbles given the same overall plastic density.
The initial internal pressure (Pi) of the bubbles is assumed
to be the same in the foam prior to incorporation into the
base mix, given a particular surfactant type and foam genera-
tor pressure. Once the foam is mixed with the cementitious
matrix, bubbles change size and the internal pressure varies in
order to maintain equilibrium with the surrounding matrix.
It is assumed that this process is elastic (Prins, 2006) and hence
small bubbles have a higher internal pressure than larger
bubbles.
This gives a coherent explanation for the observed bubble size
characteristics for different densities, water/cement (w/c) ratios,
cement and filler types. Figure 4 provides a schematic illus-
tration of the end result of these force equilibration processes.
However, this explanation does not predict instability, and thus
additional time-dependent changes to these forces must occur
in fresh mixes.
3
Magazine of Concrete Research
Freshly produced foam with
similarly sized bubbles
(the actual size is unknown)
Higher
As the bubbles are highly confined, Fc density
is high such that the pressure difference
inside and outside of the bubbles is not large.
Therefore, the bubbles do not tend to increase
in size significantly in order to achieve
equilibrium with the surrounding matrix
e.g. 600 kg/m3
Average bubble size 0·4 mm
Figure 4. Change in bubble size upon mixing foam with base
mix
Causes of stability and instability in mixes of
conventional and ultra-low densities
What is clear from observations of unstable mixes is that, at
some point in time, the bubble size becomes sufficiently large
to cause them to be buoyant and separate from the mix. The
following discussion attempts to describe the time-dependent
mechanics of buoyant unstable bubbles, and the comparative
rate at which this occurs in ‘conventional’ and ultra-low density
foamed concrete mixes.
A major time-dependent force is due to the effect of the surfac-
tant; that is, its control of the surface tension of an aqueous
bubble (Fst) (Myers, 1992; Weaire and Hutzler, 1999). For
liquid foams, the time-dependent effect of surface tension redu-
cing and leading to liquid drainage due to the effects of gravity
(Fd) is well understood. As a result, the aqueous/surfactant
liquid fraction of the foam changes, and so does the sur-
face tension of the bubbles. In turn, to maintain equilibrium,
the bubble size increases (Myers, 1992; Stevenson, 2012;
Weaire and Hutzler, 1999). However, unlike liquid foams,
bubbles in a cementitious matrix are separated by the paste or
mortar phase surrounding them. In this case, the drainage that
occurs through thin films separating the bubbles and plateau
borders (i.e. channels formed where three neighbouring films
meet (Stevenson, 2012)) in liquid foams may change. As a
result, it is not possible to directly compare the situation to
liquid foams. Stevenson (2012) reported a slower drainage rate
in foams with smaller bubbles, suggesting that drainage occurs
at a faster rate in lower density foamed concretes than those
with high densities. Furthermore, surface charges on the
bubbles and cement particles have been reported to affect the
mix stability (Jones and McCarthy, 2006). Cement particles are
4
Jones, Ozlutas and Zheng
Stability and instability of foamed concrete
Ultra-low
density
Decreased confinement and lower Fc , results in
larger pressure difference between inside
and outside of the bubbles, resulting in more
significant increases in bubble diameter to
reach equilibrium
e.g. 300 kg/m3
Average bubble size 0·6 mm
attracted to bubbles, making it more difficult for the liquid to
drain.
As already mentioned, when the mix plastic density is
decreased down to ultra-low levels, the total solids content is
decreased through reduction and eventual elimination of the
sand/filler (below 600 kg/m3), and then the cement and water
contents have to be reduced. Figure 4 represents an idealised
system in which the bubbles are considered to be of uniform
size. However, in reality, there are inevitably a range of actual
bubble sizes in a mix, each of which has a slightly different
internal pressure.
Varying bubble sizes within a mix give rise to an internal
pressure gradient. In turn, this can result in gas diffusion,
which is referred to as Ostwald ripening in liquid foams.
For aqueous foams, this is driven by the Laplace pressure
(Stevenson, 2012; Weaire and Hutzler, 1999), which is 2γ/r for
spheres, γ being the interfacial surface tension and r the bubble
radius. In this paper, γ is the surface tension. Due to this differ-
ential pressure, the gas contained in smaller bubbles diffuses
into larger bubbles, which further increase in size and further
increase the pressure differential. The process continues until
the bubbles are sufficiently buoyant to float to the surface
and burst, releasing the gas contained to the environment.
Ultimately, all foamed concretes are destroyed by this process.
In mixes where the bubbles are closer together and inter-
bubble walls are thinner (i.e. lower density mixes), this process
is easier and the process happens more quickly, as shown in
Figure 5. The SEM micrographs of low-density foamed con-
crete mixes shown in Figure 2 show the increased presence
of ‘small’ bubbles within the inter-bubble separating walls.
Magazine of Concrete Research Jones, Ozlutas and Zheng
Stability and instability of foamed concrete

F b1 F b2
Fc
diffusion
Gas

Fc

Gas diffusion Bubble has F b1 < F b2

from small to insufficient F b2 Fc
large bubble buoyancy to float
Stable mix
(a) Conventional foamed concrete
(e.g. 600 kg/m3)

Fb1 Fc
Fb2 Total
collapse

Fc Solids settled
at the bottom

Bubble becomes Segregation Instability

Gas diffusion F b1 < F b2 buoyant and floats to
from small to occurs
F b2 > F c the surface
large bubble
(b) Ultra-low density foamed concrete
(e.g. 300 kg/m3)

Figure 5. Schematic illustration of inter-bubble gas diffusion and

onset of instability in (a) ‘conventional’ and (b) ultra-low density
foamed concrete mixes based on observations

It is speculated that these could further aid inter-bubble gas
transfer and hence reduce the time for bubbles to become
buoyant.
As a result of an increase in bubble diameter (Φ), the bubble
buoyancy force (Fb) increases. Once Fb is high enough to over-
come the surrounding Fc, bubbles rise towards the surface
of the mix, displacing the surrounding solids and eventually
reaching the surface (Figure 1(b)) and causing instability
(Figure 1(c)). This dynamic environment exists until equili-
brium is reached or the mix hardens. When the mix hardens,
no further changes to the bubbles can occur. However, once
the non-equilibrium state (Fb > Fc) is reached, the process of
phase separation is irreversible.
Production of stable ultra-low density foamed
concrete
The foregoing discussion, if correct, indicates that the only
way to prevent instability is for the mix to ‘solidify’ prior
to bubbles becoming large enough to become buoyant. For
denser foamed concretes this is easily achieved within the
typical initial set times of Portland cement (PC). However,
for ultra-low density foamed concrete this is not fast enough.
The authors experimented in the laboratory with a range
of high early strength PCs, increased mix temperatures
and accelerating admixtures, but none were found to be
consistently successful. Thus, further laboratory trials were
undertaken using a blend of PC with a compatible calcium
sulfoaluminate (CSA) cement, the results of which are now
described.
Constituent materials, mix proportions and
production of foamed concrete
The following constituent materials were used to produce
foamed concrete mixes for testing.
& CEM I 52.5N (PC) conforming to BS EN 197-1 (BSI,
2011).
& Commercial CSA cement compatible with PC as an
additive to provide rapid setting.
& Fine aggregate (natural siliceous sand) conforming to
BS EN 12620 (BSI, 2008) category GF85.
& Surfactant (a commercially available protein-based foaming
agent), used in a 6% aqueous solution and foamed to a
density of 50 ± 5 kg/m3.
The methods used for designing, producing and curing the
foamed concrete were as described by Jones and McCarthy
(2005), except that a tolerance of ±25 kg/m3 of the target
plastic density was used rather than the more typical
±50 kg/m3, as the latter could represent 25–50% of the target
plastic density.

5
Magazine of Concrete Research Jones, Ozlutas and Zheng
Stability and instability of foamed concrete

Test methodologies
Setting time
The initial setting time of the paste fraction was measured
with an automatic Vicat apparatus in accordance with BS EN
196-3 (BSI, 2005) using a w/c ratio of 0·5.
Stability
Stability was measured by pouring fresh foamed concrete
mixes into 500 mm deep and 75 mm diameter polycarbonate
cylinders (see Figure 1(a)) lined with polythene film. The mix
was further observed over 24 h for any reduction in height, to
measure any longer term instability.
Bubble size analysis
Bubble size analysis was carried out using optical microscopy
and automated image analysis software. Test samples were
obtained from 500 mm high cylindrical specimens after 28 d of
sealed curing. The cylinders were split longitudinally and then
sections from the top, middle and bottom (in the direction of
cast) of the cylinder were taken and the average of these used
to give mean bubble diameters. Broken surfaces were cleaned
of dust and sprayed with fluorescent paint to improve image
contrast under UV illumination. A microscope-mounted
digital camera was used to capture a 100 mm2 image (with res-
olution of ≈2000  2000 pixels), and 2D image analysis was
carried out using ImageJ software (Ferreira and Rasband,
2012), using a similar approach to that described elsewhere
(Nambiar and Ramamurthy, 2007; Visagie and Kearsley, 1999).
Results and discussion
Stability
For mixes with plastic densities from 1000 kg/m3 down to
400 kg/m3, 100% PC consistently produced stable foamed con-
cretes. Then, on a trial-and-error basis, lower density mixes
were made stable by incorporating CSA cement to partially
replace PC. Firstly, 5% of PC (by mass of cement) was
replaced with CSA, producing stable 300 kg/m3 and 200 kg/m3
foamed concrete mixes. The CSA content was increased to
10% (by mass of PC) to produce stable 150 kg/m3 density
mixes consistently, as summarised in Table 1. Table 2 shows
the relationship between collapse and base mix setting times.

Cement: kg/m3
Plastic density: CSA: % by mass Water: Sand: Air volume:
kg/m3 CEM I CSA of cement kg/m3 kg/m3 %

1000 300 — — 150 550 55

600 300 — — 150 150 70
500 335a — — 165 — 73
400 267 — — 133 — 78
300 190 10 5 100 — 84
200 126 7 5 66 — 89
150 90 10 10 50 — 92

a
Cement content increased to increase ‘fines’ content as sand was not used below 600 kg/m3

Table 1. Test mix constituent proportions; a w/c ratio of 0·5 was

used for all mixes

Base mix
CEM I: CSA: initial setting Foamed concrete collapse time
% by mass % by mass time: hh:mm in the absence of CSA: hh:mm Stability: Stable (S) or unstable (U)

D150a D200 D300 D150 D200 D300 >D300

100 — 03:25 01:10 01:40 01:55 U U U S/U
95 5 01:30 n/mb n/ac n/a U S/U S S
90 10 00:20 n/a n/a n/a S S S S

b
D is plastic density value in kg/m3
c
Not measured
Not applicable

Table 2. Collapse and initial setting times; w/c ratio = 0·5

6
Magazine of Concrete Research
95% PC/5% CSA
1·0
Average bubble diameter: mm
0·8
0·6
0·4
0·2
0 0
200
Figure 6. Bubble size characteristics of stable foamed concrete
mixes
It was not possible to produce mixes with plastic densities
below 150 kg/m3 as the CSA content had to be increased
above 10%. This caused the base mix to set within 2 min and
there was insufficient time to incorporate foam and place the
foamed concrete. However, with the use of CSA set controllers,
producing stable foamed concrete below 150 kg/m3 could be
possible.
Bubble size analysis
Figure 6 summarises the results of bubble size analysis carried
out by 2D image analysis in relation to stability. As expected,
bubbles increased in diameter in the ultra-low density foamed
concretes. The average bubble diameter increased 2·6 times
from the highest to lowest density. The 150 kg/m3 density
samples did not split cleanly but fractured into multiple pieces,
and thus it was not possible to carry out analysis on these
samples.
Figure 6 also gives the calculated bubble to solid area ratio
obtained from the 2D image analysis. As the plastic density
decreased this ratio increased due to the increased air content
at 300 kg/m3 density, where stability issues commence with util-
isation of 100% PC, the bubble to solid area is around 1. As
discussed earlier, the buoyancy force of the bubble (Fb) will
then tend to be similar to the ‘confinement’ force (Fc) and
such mixes will be on the borderline between being stable or
becoming unstable with typical PC set times.
Conclusions
By considering the internal forces likely to be affecting bubbles
and the surrounding paste/mortar fractions of foamed con-
crete, it is possible to present a coherent reasoning and mech-
anism to explain instability in fresh foamed concrete mixes.
Jones, Ozlutas and Zheng
Stability and instability of foamed concrete
2·0
100% PC 100% PC (with sand)
1·8
Average bubble diameter 1·6
Bubble/solid area ratio
Bubble/solid area ratio 1·4
Larger area of 1·2
bubbles to solid
1·0
Larger area of
solids to bubbles 0·8
0·6
0·4
0·2
300 500 600 1000
Plastic density: kg/m³
The underlying cause of instability is considered to be the
buoyancy force of the bubbles, which allows them to float out
of a fresh mix and ultimately cause complete separation of the
gas and solid phases. The buoyancy force is directly related to
bubble size, and this becomes significantly larger at lower den-
sities; larger bubbles are consequently much more buoyant and
hence lower density mixes are more prone to instability.
This is explained by the observation that, as the air content
fraction is increased in lower density mixes, the ‘confinement’
force due to the solids is reduced. In addition, larger bubbles
and a smaller solids fraction result in bubbles being closer and
the separating walls thinner. Established theory predicts that
this makes it easier for gas to diffuse from smaller (high
internal pressure) to larger (low internal pressure) bubbles, and
hence the bubbles become more buoyant over time. At a criti-
cal time, related to the bubble buoyancy, separation of solid
and gas phases (i.e. instability) occurs and the fresh mix col-
lapses. Following extensive laboratory trials, the best method
of overcoming this was found to be the blending of rapid-
setting CSA cement with the PC.
The practical issues of placing ultra-low density foamed con-
cretes with base mixes produced from cement with an initial
setting time of 20–25 min are problematic. Further research is
needed to develop methods to retard the initial setting times
for transportation and mixing, and then activate it immediately
when rapid hardening is needed after placement.
Acknowledgements
The authors wish to acknowledge Propump Engineering Ltd
for help and discussions. Thanks also go to CALTRA
Nederland BV for their support with the supply of CSA.
7
Magazine of Concrete Research
REFERENCES
Aldridge D (2005) Introduction to foamed concrete: what, why,
how? In Use of Foamed Concrete in Construction (Dhir RK,
Newlands MD and McCarthy A (eds)). Thomas Telford,
London, UK, pp. 1–14, http://dx.doi.org/10.1680/uofcic.
34068.
BSI (2005) BS EN 196-3:2005+A1:2008. Methods of testing
cement. Determination of setting times and soundness.
BSI, London, UK.
BSI (2011) BS EN 197-1:2011. Cement. Part 1: composition,
specifications and conformity criteria for common cements.
BSI, London, UK.
BSI (2013) BS EN 12620:2002+A1:2008. Aggregates for
concrete. BSI, London, UK.
Ferreira T and Rasband W (2012) ImageJ User Guide, IJ1.46th
edition. National Institute of Mental Health, Bethesda,
MD, USA.
Jones MR and McCarthy A (2005) Utilising unprocessed
low-lime coal fly ash in foamed concrete. Journal of Fuel
84(11): 1398–1409.
Jones MR and McCarthy A (2006) Heat of hydration in foamed
concrete: effect of mix constituents and plastic density.
Cement and Concrete Research 36(6): 1032–1041.
Jones MR and Zheng L (2013) Energy absorption of foamed
concrete from low–velocity impacts. Magazine of Concrete
Jones, Ozlutas and Zheng
Stability and instability of foamed concrete
Research 65(4): 209–219, http://dx.doi.org/10.1680/macr.
12.00054.
Myers D (1992) Surfactant Science and Technology, 2nd edn,.
VCH Publishers, Inc., New York, NY, USA.
Nambiar EKK and Ramamurthy K (2007) Air-void
characterisation of foam concrete. Cement and Concrete
Composites 37(2): 221–230.
Prins A (2006) Surface forces’ role in the foaming
behaviour of liquids. In Encyclopedia of
Colloid and Surface Science, 2nd edn, vol. 8
(Somasundaran P (ed.)). Taylor & Francis, Abingdon,
UK, pp. 5926–5937.
She W, Yunsheng Z and Jones MR (2014) Using the ultrasonic
wave transmission method to study the setting behaviour of
foamed concrete. Construction and Building Materials 51:
62–74.
Stevenson P (2012) Foam Engineering: Fundamentals and
Applications. Wiley, Chichester, UK, pp. 75–89.
Visagie M and Kearsley EP (1999) Micro-properties of foamed
concrete. In Specialist Techniques and Materials for
Concrete Construction (Dhir RK and Henderson NA
(eds)). Thomas Telford, London, UK, pp. 173–184,
http://dx.doi.org/10.1680/stamfcc.28258.
Weaire D and Hutzler S (1999) The Physics of Foams.
Clarendon Press, Oxford, UK.
WHAT DO YOU THINK?
To discuss this paper, please submit up to 500 words to
the editor at journals@ice.org.uk. Your contribution will
be forwarded to the author(s) for a reply and, if con-
sidered appropriate by the editorial panel, will be pub-
lished as a discussion in a future issue of the journal.